What is it about?
Grid-based hex meshing is robust, but its axis-aligned boundary converges to the input geometry at only first order, and the heuristic padding-and-projection steps offer no quality guarantees. MCHex instead extracts a marching cubes surface as the mesh boundary and produce an all-hex mesh. It guarantees a positive Jacobian for every element, naturally padded boundaries, and second-order boundary convergence. On 202 benchmark geometries, MCHex runs the fastest among existing methods, using fewer elements to achieve similar boundary fidelity.
Featured Image
Photo by Buddha Elemental 3D on Unsplash
Why is it important?
MCHex is the first grid-based hex meshing method with a rigorous positive-Jacobian guarantee for every element. It runs the fastest among existing methods, using fewer elements to achieve similar boundary fidelity. This makes hex meshing---long an open problem blocking reliable finite element analysis on complex geometries---both robust and fast.
Perspectives
Hex meshing progress has been slow for decades, largely because the final projection step has remained heuristic. I have been working on this open problem for two years. I believe that providing methods with quality guarantees and convergence-rate guarantees is important, and I hope to see more research in this direction.
Hua Tong
Carnegie Mellon University
Read the Original
This page is a summary of: MCHex: marching cubes based adaptive hexahedral mesh generation with guaranteed positive Jacobian, ACM Transactions on Graphics, October 2026, ACM (Association for Computing Machinery),
DOI: 10.1145/3858338.
You can read the full text:
Contributors
The following have contributed to this page
